The Aging Microbiota-Gut-Brain Axis in Stroke Risk and Outcome

Antibiotics:

Although a useful tool to manipulate the host gut microbiota, antibiotic administration can have off-target effects on microbiota at other sites (e.g., lungs, skin, eyes, oropharynx, etc.). Some antibiotics can also elicit microbiota-independent changes in host metabolites and phagocytic clearance of pathogens by reducing cellular respiration in immune cells²³. Antibiotics can alter cellular metabolism and lead to hyperglycemia, increased adiposity, and insulin resistance, known risk factors for stroke²⁴. Colonocytes adapted their preference for SCFAs as an energy source toward a preference for glucose via increased glucagon-like peptide 1 (GLP-1) signaling in response to a sudden reduction of luminal SCFAs following antibiotic treatment²⁵. The effects of antibiotics depend on the antimicrobial coverage, dose, duration, and host factors, such as age, sex, diet, strain, and presence of other pathologies²⁶, which must be carefully validated for the specific experimental conditions.

Fecal Microbiota Transplant (FMT):

Successful FMT requires a diligent design and validation of several parameters including age, sex, strain, vendor, disease, and treatment status of both donors and recipients. Collection methods, time-to-freeze versus fresh use, freeze-thaw cycles, oxygen exposure, and preservatives must be evaluated and reported. An important clinical and pre-clinical consideration in FMT use is that all microbial content (bacterial and non-bacterial) is transferred. Slowly degradable metabolites, viral, fungal, and parasitic genes are also introduced into the recipient. These non-bacterial entities pose a clinical risk of transferring infections and can be a source of experimental confounders.

In Vitro Systems such the gut-on-a-chip²⁷ or SHIME (Simulator of the Human Intestinal Microbial Ecosystem²⁸) are used to investigate the direct implication of the microbiota on gut homeostasis. Purified intestinal bacterial cultures derived from a single healthy donor’s stool have been used to treat recurrent hypervirulent C. difficile infection in patients who had failed antibiotic treatments²⁹. In the same study, recipients remained symptom-free at six months after FMT and their stool samples demonstrated over 25% similarity to the synthetic stool substitute²⁹. In vitro approaches may effectively address the FMT’s clinical safety concerns³⁰. Human intestinal organoids grown from induced pluripotent or embryonic stem cells are used to study clinical pathologies such as response to C. difficile toxin³¹. These organoids are sterile (unless bacteria are introduced), have a functional epithelial barrier, and are composed of lysozyme- and mucin-producing cells that resemble Paneth and goblet cells (see Table 1)³². Organoid models provide experimental control over the oxygen and nutrient levels. To accelerate drug development, engineered microchips containing living human cells that reconstitute organ-level functions have been developed^33,34.

Gnotobiotic Models:

Germ free (GF, devoid of detectible bacteria) or gnotophoric (one or a few known bacteria) models have revealed microbiota’s role in the development of the immune and nervous systems. GF models have limitations, such as defective immune responses³⁵, reduced blood-brain barrier integrity, and altered microglial maturation and function, and therefore may be less translationally relevant for human studies. Specific changes observed in GF mice are summarized in Table 2. The artificial nature of animals raised in vivaria may also lead to conclusions with less translational relevance. Other model organisms, including studies in Drosophila and Zebrafish have explored the MGBA and have been recently reviewed³⁶. There are also differences between breeders and mouse strains that impact MGBA signaling³⁷. Importantly, baseline differences in MGBA immunity, neurochemistry, and behavior in GF mice are potential experimental confounders in preclinical studies of the role of MGBA in stroke.

Enormous advances in the tools available to study the role of MGBA in stroke have been made. However, potential pitfalls make interpretation of existing evidence and design of future studies challenging. These pitfalls include differences in the anatomy, immunology, microbiota composition, multi-comorbidities including aging, sample preparation, bioinformatics approaches, and thresholds for distinguishing “correlation versus causation” in MGBA research³⁸. For instance, the role of gut virome or bacteriophages of the microbiome in stroke have not been investigated. Attempts for selective bacterial colonization remain artificial, neglecting the effects of complex microbial signaling such as quorum sensing and biofilm³⁹. Despite inherent limitations, gnotobiotic models remain valuable in establishing causality in MGBA studies of stroke, discussed later in this review.

Immune Surveillance In The LP:

Immune tolerance is defined as the absence of an effector response to an immuno-potent antigen⁷¹. In addition to M cells (Table 1), tolerogenic (CD103+) DCs directly sample luminal antigens (Figure 1) and descend from the epithelium to present their antigen to naïve T cells, inducing their differentiation into anti-inflammatory T_reg cells. T_reg cells travel to the LP to secrete anti-inflammatory cytokines (e.g., IL-10 and TGF-β) to maintain immune quiescence⁷¹. Gut microbiota regulate the mucus layer and antimicrobial function of goblet and Paneth cells (Table 1), and plays an active role in inducing LP immune cells, particularly LP lymphocytes (Table 2). The role of microbiota in LP immunity and inflammation has been extensively reviewed elsewhere⁷².

CNS Myeloid Cells:

The gut microbiota influences microglia (MG) biology throughout life (Tables 1 and 2). Within minutes after injury, MG sense DAMPs (e.g., ADP and ATP via their P2Y12 receptor) to initiate chemotaxis to the site of injury. Upon activation, MG retract their processes to transform into an amoeboid morphology and begin coordinating with other CNS cells and peripherally-sourced immune cells. Microbial metabolites directly regulate MG-mediated neuroinflammation^62,73. Tryptophan-derived indole-based ligands activate the MG aryl hydrocarbon receptor (AHR, a regulator of immune differentiation and neuroinflammation) to suppress the activation of the NF-κB pathway. Microbiota-dependent activation of AHR reduces VEGF-β expression by MG and increases TGF-α secretion by astrocytes to dampen astrocyte-mediated neuroinflammation through a SOCS2-dependent pathway (Figure 3a)⁷⁴. CNS BAMs, which line the meninges, choroid plexus, and perivascular spaces (Figure 1), are CNS-resident myeloid cells, distinguished from MG by higher expression of CD38 and MHC-II, while lacking typical MG markers (Tmem119 and P2RY12)^75,76. The choroid plexus is a site high cerebral blood flow and a site of recruitment for inflammation-resolving immune cells from the peripheral circulation^77–79. At blood-CSF interfaces such as the choroid plexus and meninges, BAMs and other APCs continuously monitor and filter brain-derived antigens at steady state and after stroke.

Cytokines:

The gut microbiota can drive CNS pathology both through its metabolites and by modulation of circulatory cytokines. Dysbiosis can lead to increased circulatory proinflammatory cytokines such as IL-1β, IL-6, TNF-α, and IFN-γ, disruption of tight junctions and increased permeability across the intestinal epithelial barrier, GBB, and BBB⁹⁷. Dysbiosis-induced disruption of these barriers provides a direct bidirectional gateway for the brain-derived antigens to trigger a systemic immune response and for microbiota-derived metabolites to access the CNS immune compartment. It is difficult to discern the direct effects of metabolites or cytokines on the CNS from their indirect effects via immunological and peripheral neural pathways⁹⁸. Bile acids, SCFAs, and tryptophan metabolites represent well-studied classes of microbiota metabolites that mediate the MGBA communications and will be discussed under the MGBA and Stroke section of this review.

Microbiota And Age-Related Changes In Innate Immunity:

Consistent with reported increased proliferation and tissue-redistribution of myeloid cells with aging, we have shown that peripherally-sourced APCs significantly increase in the brain with advancing age¹¹⁴. This myelocytic accumulation in the brain positively correlated with age-related changes in the gut microbiota (decreased Akkermansia) and behavioral deficits¹¹⁴. Although the proportion of neutrophils in circulation does not appear to be affected, their functions (chemotaxis, phagocytosis, production of ROS) are impaired with aging^101,115. The microbiota drives neutrophil aging via TLR and MyD88 signaling, and depletion of the microbiota (in GF or antibiotic-treated mice) reduced the number of circulating aged neutrophils (CXCR4^highCD62L^low) and improved survival in animal models of endotoxin-induced septic shock¹¹⁶.

Microbiota And Age-Related Changes In Adaptive Immunity:

Age-related changes in T cell populations include decreased numbers of naïve T and T_reg cells and increased CD8+ memory and Th17 cells¹¹⁷. Age-related changes in the B cell include a decline in B cell lineage precursors¹¹⁷. With aging, peripheral B cell phenotypes shift away from naïve B cells toward antigen-experienced B cells with restricted clonal diversity with impaired capacity to produce novel high-affinity immunoglobulins¹¹⁸. Age-associated subsets of B cells have been described as a population with high expression of CD11c and the transcription factor T-bet. These B cells accumulate in the blood and meninges with age and can undergo an exaggerated proliferative response after stimulation with TLR7 and TLR9 ligands^119,120.

Studies have shown that the marked shift from lymphocyte proliferation to myeloid cell proliferation with aging in SPF mice is abrogated in GF mice¹²¹. There is an expansion of B cells in GF mice, suggesting a suppressive role of microbiota on B cell proliferation in the bone marrow. Additionally, A. muciniphila supplementation in an accelerated aging model (Ercc1−/Δ7 mice with impaired DNA repair protein ERCC1) decreased the number of activated CD80+CD273− B cells in PPs¹⁰⁵. These findings implicate the gut microbiota in the regulation of lymphocyte proliferation and suggest a possible treatment opportunity for lymphocyte-dominant disorders (e.g., autoimmune diseases) and as a potential adjunctive therapy used in combination with approved B cell depletion medications.

Aging And CNS Immunity:

Age-related changes in CNS immune cells are summarized in Table 1. Microbiota depletion using antibiotics or in GF models have demonstrated a detrimental role for the microbiota in AD¹²². FMT from WT controls into APPswe/PS1dE9 transgenic mice (a model of AD) improved cognitive deficits, reduced brain deposition of amyloid-β (Aβ), and decreased neuroinflammation (reduced COX-2 and CD11b expression) in the brain. These effects were associated with increased fecal SCFAs (butyrate)¹²³. GF AD mouse models showed reduced Aβ deposition. Supplementation of SCFAs in GF AD mice upregulated MG expression of the ApoE gene and increased Aβ deposition, suggesting that microbiota-derived SCFAs promote Aβ deposition, likely by regulating MG phenotype and phagocytosis¹²⁴. Lymphatic drainage of CNS antigens is decreased with aging^125,126, suggesting that exposure of brain-derived antigens to BAMs and other APCs may be prolonged in aged compared to young stroke. FMT from aged donors into young recipients increases MG activation (decreased P2RY12 and increased Iba-1 and MHCII expression) and systemic neuroinflammation^127,128. Moreover, GF mice that received FMT from aged WT mice show decreased fecal SCFAs (acetate, propionate, and butyrate) and demonstrate depressive-like behavior, impaired short-term memory, and impaired spatial memory over the 3 months following FMT¹²⁹.

Dysbiosis has emerged as a therapeutically-targetable contributor to immunosenescence¹³⁰. Rejuvenation procedures (co-housing, serum-injections, and heterochronic parabiosis) showed that the microbial community and intestinal immunity of aged mice becomes comparable to those of young mice¹³¹. Metagenomic data implicated that a higher abundance of Akkermansia and butyrate-producers were responsible for the immune rejuvenation seen in aged mice¹³¹. These findings suggest that microbiota can drive age-associated neuroinflammation, independently of the host’s age, and can contribute to post-stroke inflammation^127,132.

Microbiota And Age-Related Changes In Sympathetics/Paraympathetics:

Age-related changes in the ANS function include decreased neurotransmitter function, leading to diminished regulatory activity (e.g., CBF regulation) and poorly coordinated autonomic discharge (e.g., bladder function)¹³⁷. Resting sympathetic tone increases with aging, which is correlated with increased plasma NE and decreased sensitivity of the catecholamine and cholinergic receptors¹³⁷, which implies an age-related cholinergic dysfunction of CAIP after stroke (discussed above). FMT from spontaneously hypertensive rats into normotensive rats increases sympathetic activity, blood pressure, and neuroinflammation¹³⁸, suggesting that microbiota can mediate sympathetically-driven pathophysiology seen with aging and in stroke.

Microbiota And Age-Related Changes In ENS:

ENS neurons undergo age-related neurodegeneration in animals¹³⁹. The loss of enteric neurons with aging reduces the density of nerve fibers in the intestinal mucosa¹⁴⁰. Aged enteric neurons are swollen and dystrophic and exhibit lipofuscin accumulation and higher ROS production in the myenteric plexus¹⁴¹. Age-related reductions in EGCs, proportional to loss of enteric neurons, has also been described¹⁴². Although, communication pathways are present between the microbiota and ENS, the direct contribution of age-related dysbiosis to ENS neurodegeneration has not been elucidated.

In summary, all components of the MGBA undergo significant age-related alterations, many of which can be directly influenced by or even driven by the microbiota. Given these observations and the fact that stroke is a disease of the aged, studies of stroke pathophysiology and outcomes must consider the age-associated background in the host and the microbiota, including dysbiosis, GI dysfunction, immune remodeling, and other existing CNS or systemic pathology.

MG Response After Stroke In Aged Animals:

Complete depletion of MG (using the dual CSF1R/c-Kit inhibitor PLX3397) worsens stroke outcomes¹⁷⁴, suggesting that MG play a net beneficial role after stroke, likely by suppressing the post-stroke inflammation mediated by injured neuroglia. Aging affects the MG response to stroke^101,172,175. Aged MG produce less TNF-α, but more ROS than young MG, and aged mice reconstituted with young bone marrow had reduced MG activation (lower FGF production and overall granularity) after stroke¹⁷². Additionally, post-stroke FMT from young healthy mice into aged mice reduced MG activation (decreased Iba1 and increased P2RY12)¹²⁸. These results indicate that the age-dependent role of MG after stroke can be influenced by the manipulation of gut microbiota.

Innate Immune Response After Aged Stroke:

Post-stroke neutrophil infiltration into the brain is increased in aged animals¹⁷². Aged neutrophils show reduced phagocytosis but produce higher levels of ROS and MMP-9 after storke^115,172. CNS BAMs, DCs and blood-derived monocytes are also among the early post-stroke responders. CCR2-mediated recruitment of monocyte-derived macrophages can be protective in the early phases of stroke by contributing to phagocytic activity and debris clearance; however, their continued presence contributes to chronic inflammation. Spleen and bone marrow are the primary sources of monocytes¹⁷⁶ and contribute to systemic inflammation, including gut inflammation. In fact, upregulation of TLR4 expression on peripheral blood monocytes from acute stroke patients positively correlates with higher stroke severity and inflammatory injury¹⁷⁷, implicating these monocytes in stroke-induced systemic inflammation. It is unclear whether age-related alterations in the microbiota can alter the post-stroke involvement of peripherally sourced neutrophils and monocytes.

DCs arrive early and are detected for several days after the onset of stroke¹⁷⁸. Under the microbiota’s control, DCs contribute to inflammatory signaling by secreting IL-23, inducing the expression of IL-17 by γδ T cells, and promoting neutrophil infiltration¹⁷⁹. The pre-stroke number of brain APCs is significantly higher in aged mice, and is correlated with specific age-related shifts in the microbiota (e.g., abundance of Akkermansia is negatively correlated with number of brain APCs)¹¹⁴. The post-stroke composition of infiltrating myeloid cells shows a larger monocytic contribution migrating into the young brain, but a larger neutrophilic contribution infiltrating into the aged brain 3 days after stroke¹⁷². The skull bone marrow houses significant populations of monocytes and neutrophils, with transcriptional signatures distinct from their blood-derived counterparts¹⁸⁰. Importantly, the composition of myeloid populations in the skull bone marrow changes with aging with significantly higher neutrophils in the aged skull (Figure 4a, original unpublished data), which may explain the increased neutrophilic infiltration after stroke in aged compared to young mice.

The Adaptive Immune Response After Stroke In Aging:

The subacute response begins 1 to 3 days after stroke. As a result of ischemic injury and local inflammation, critical interfaces including the BBB, meningeal lymphatics, and blood-CSF barriers become compromised^169,181, allowing brain-derived antigens and cytokines to leak into the systemic circulation. With help from APCs (primarily DCs), the adaptive arm is engaged for an antigen-specific response and development of immune memory for the presented brain antigens. B and T lymphocytes undergo clonal expansions in lymphoid organs (e.g., spleen and cervical lymph nodes) and return to the circulation to mount a whole-body response. Increased circulating lymphocyte subsets (e.g., CD4+CD28− beyond 8%) have been linked to a higher risk of recurrence and mortality after stroke¹⁸². Increased IL-17+ T cells in the blood is associated with exacerbated cognitive deficits after stroke^183,184, suggesting a contribution of antigenic activation of lymphocytes to chronic sequelae of stroke such as cognitive decline and autoimmunity. Indeed, autoreactive B and T cells are detected as early as four days after stroke, and circulating lymphocytes from stroke patients collected two weeks after their stroke show a more robust immunoreactivity to brain antigens (e.g., myelin) than those collected from controls¹⁸⁵.

Pre-stroke bacterial colonization (by co-housing) from non-stroke SPF mice into GF mice was neuroprotective (reduced infarct volumes and increased IL-1β and TNF-α) in GF mice. This beneficial effect of microbiota was abolished in lymphocyte-deficient GF (GF Rag1−/−) mice, indicating that microbiota from healthy donors leads to lymphocyte-mediated neuroprotection after stroke¹⁸⁶. Recent reports indicate that meningeal B cells, derived locally from the skull bone marrow, provide a constant supply of B cells exposed to brain antigens¹²⁰. Additionally, we have found that skull bone marrow contains significantly higher CD8+ T lymphocytes and activated CD11b^high B lymphocytes in naïve aged mice when compared to young skull (Figure 4b, original unpublished data). Blood-derived antigen-experienced B cells populate the meninges in aged mice, but not in young mice, in the absence of an acute CNS injury¹²⁰.

Future studies are needed to determine whether these age-dependent heterogeneity of myeloid cells and lymphocytes contribute differently to the post-stroke response. Their proximity, temporal engagement, and cell-specific regulation by the gut microbiota after acute brain ischemia combined with the slower lymphatic clearance seen with aging¹⁸⁷ may result in a prolonged exposure to novel CNS antigens that could contribute to increased incidence of autoimmunity and long-term cognitive decline seen in stroke patients^188,189.

These findings indicate that aging alters the innate and adaptive immune response after stroke with an increased neutrophilic infiltration, an earlier involvement of peripherally-sourced APCs, and potentially different contribution of skull bone marrow-derived immune cells compared to immune cell sourced from the peripheral lymphoid organs (e.g, peripheral bone marrow, spleen, and thymus). It can be speculated that the age-related decrease in protective bacteria (e.g., Akkermansia or SCFA-producers) can lead to reduced intestinal barrier function. This then activates DCs to stimulate γδ T cells, increasing their migration to the brain and enhancing the production of IL-17, leading to the additional recruitment of neutrophils from the blood or skull bone marrow into the aged brain after stroke (Figure 3b). Future experiments are needed to clarify the role of age-related changes in the gut microbiota in cell-specific immune responses after stroke and in long term outcomes of stroke such as cognitive decline.

Clinical Studies Of The Role Of MGBA After Stroke:

Table 4 provides an up-to-date summary of clinical studies focused on the role of MGBA in stroke. Studies have found a decrease in bacterial diversity in samples from stroke patients compared to healthy controls^191,192. Yamashiro et al.¹⁹³ found that increased Lactobacillus ruminis and reduced SCFAs (acetate and valerate) positively correlated with post-stroke inflammatory markers (elevated C-reactive protein and increased white blood cell counts) in a Japanese cohort of stroke patients. Xia et al.¹⁹⁰ reported that Parabacteroides, Oscillospira, and Enterobacteriaceae were enriched in stroke patients, while Prevotella, Roseburia, and Faecalibacterium were reduced. Xu et al.¹⁹⁴ showed that a higher abundance of Enterobacteriaceae was positively correlated with poor outcomes. A well-defined disease-specific microbial “fingerprint” for post-stroke dysbiosis requires large patient cohorts with appropriate controls for comorbidities, geography, diet, medication, race, sex, and age.

Pre-Clinical Studies Of The Role Of MGBA After Stroke:

Table 5 provides an up-to-date summary of pre-clinical studies focused on the role of MGBA in stroke, separated into studies that used young animal models versus those that used aged animal models of experimental stroke. In pre-clinical models, a complete absence of the microbiota in GF mice and antibiotic-treated mice leads to poorer stroke outcomes compared to SPF mice¹⁹⁵, which suggest a net beneficial role of young microbiota. Moreover, pre-stroke bacterial colonization from non-stroke SPF mice into GF mice reduces infarct volumes, increases MG/macrophage counts, and increases brain cytokines (IL-1β and TNF-α) in GF mice, indicating that the pre-stroke microbiota from young healthy donors has the potential to play a neuroprotective role after stroke¹⁸⁶. On the other hand, pre-stroke FMT from SPF mice with stroke increased infarct volume and worsened neurological deficits after stroke in GF mice¹⁹⁶. Consistently, pre-stroke FMT from stroke patients into antibiotic-treated mice increased infarct volume and increased neurological deficits in recipient mice¹⁹⁰. These results suggest that pre-stroke colonization with microbiota from healthy donors is neuroprotective, but colonization with a dysbiotic microbiota (as seen with aging, and in patients with prior stroke or other neurological disease) has deleterious effects on stroke outcomes. The effects of a prior stroke on the immune response to a new stroke is of significant clinical relevance and should be further investigated.

Importantly, aged mice not only have a baseline dysbiotic state, they are also more susceptible to stroke-induced dysbiosis as measured by increased gut permeability and sepsis due to bacterial translocation¹⁹⁷. Aged mice receiving pre-stroke FMT from young donors have improved survival rates, reduced circulating cytokines (IL-6, TNF-α, Eotaxin, and CCL5), and improved behavioral outcomes after stroke¹²⁷, even though infarct size was unaltered. This suggests that microbiota rejuvenation in older patients at risk for stroke could be a promising prophylactic therapy. With the exception of a few studies¹²⁸, most preclinical studies have been performed in young mice and microbiota manipulations have been performed before stroke onset, both of which significantly reduce the translational value of the results (Table 5).

Probiotics And Prebiotics:

Probiotics (live bacteria) and prebiotics (selectively fermented ingredients that promote changes in the composition and activity of the microbiota) can be substitutes for clinical FMT. For instance, Panax notoginsenoside extract given before stroke is protective in rats by regulating GABA-b receptors via an increase in Bifidobacterium longum¹⁹⁸. Pre-stroke administration of probiotics can also suppress the production of proinflammatory cytokines (TNF-α and IL-6), reduces hippocampal neuronal injury, and restores spatial memory behavior in a mouse hypoperfusion-model of stroke¹⁹⁹. A recent meta-analysis showed that supplementation of enteral nutrition with probiotics in stroke patients is associated with decreased serum TNF-α, IL-6, and IL-10. Probiotics supplementation also reduced the incidence of post-stroke complications, including esophageal reflux, bloating, constipation, diarrhea, gastric retention, and gastrointestinal bleeding²⁰⁰.

Short-Chain Fatty Acids (SCFAs):

Dietary fibers and resistant starch are degraded via bacterial fermentation into acetate, propionate, butyrate, and other similar molecules (five carbons or less), known as SCFAs. SCFAs decrease intestinal inflammatory cytokines, regulate intestinal secretion, and enhance peristalsis. Although the host can produce some SCFAs (e.g., acetate²⁰¹), large quantities of SCFAs are produced by the gut microbiota. SCFAs exert their effects in multiple ways, including binding G-protein coupled receptors (GPCRs), passive diffusion through the cell membrane, stimulation of histone acetyltransferases, inhibition of histone deacetylase 3 (HDAC3), stabilization of the hypoxia-inducible factor (HIF), and have vasoactive properties^202,203. GPCRs for SCFAs (FFAR2/3) are expressed on epithelial, enteroendocrine²⁰⁴, and innate immune cells²⁰⁵, as well as on vagal afferent fibers²⁰⁶. SCFAs (butyrate) increased the production of AMPs and proinflammatory cytokines (IL-6 and TNF-α) by the intestinal epithelium, indicating a role for SCFAs in host anti-microbial defense²⁰⁷. SCFAs decreased the production of LPS-induced TNF-α, IL-1β and IL-6 by murine macrophages²⁰⁸. These findings suggest that the effects of SCFAs are cell- and pathway-specific. Further evidence is needed to determine whether physiologically relevant concentrations of SCFAs can reach the brain, given their short half-life (25 min – 3 hours)²⁰⁹.

SCFA-Producing Bacteria In Stroke:

Age-related loss of SCFA-producers has been linked to stroke risk factors, age-related cognitive decline, and stroke recovery^128,202. Fecal metagenomic analysis of patients with hypertension shows a decline in SCFA utilization, and FMT from human donors with hypertension versus normotensive controls into GF mice show transmissibility of the hypertensive phenotype via the microbiota²¹⁰. Mono-colonization of GF mice with butyrate-producing bacterium improves BBB integrity²¹¹. Our lab has shown that aged microbiota alone is sufficient to produce lower fecal SCFAs and led to enhanced cognitive deficits in GF mice, when compared to GF mice with young microbiota¹²⁹/ This indicates that age-related dysbiosis (characterized by reduced SCFAs among other features) can contribute to stroke risk factors and cognitive decline in the absence of an acute ischemic stroke.

Emerging evidence indicates that microbiota composition can be therapeutically exploited to improve recovery even when initiated after stroke in aged mice either with young FMT or any targeted manipulation of the microbiota. Post-stroke bacteriotherapy with SCFA-producers (B. longum, C. symbiosum, F. prausnitzii, and L. fermentum) and inulin (a prebiotic substrate for SCFA producers) initiated three days after stroke, 1) increased brain and plasma levels of SCFAs, 2) increased expression of mucin-related genes and improved intestinal barrier function, 3) reduced the frequency of proinflammatory IL-17-producing γδ T cells in the brain (not in the intestines), and 4) improved stroke recovery (improved neurological deficits, motor function, grip strength, and depressive phenotypes) in aged mice¹²⁸. These beneficial effects were independent of infarct size. Of significance, these effects on stroke outcomes were not related to the chronological age of the host, but instead were mediated by the age of the microbiota or its metabolites. Complementing this finding, a recent study showed that IL-17+ γδ T cells traffic from the intestinal LP to the leptomeninges and increase neuroinflammation after stroke¹⁷⁹. The beneficial effects of SCFAs after stroke may be mediated by an increase in anti-inflammatory IL-10 and a decrease in gut-originated IL-17+ γδ T cells in the ischemic brain. These results indicate that post-stroke approaches targeting the microbiota composition and function can be a promising clinical strategy with an extended therapeutic window for the treatment of stroke.

Bile Acids (BAs) are synthesized from cholesterol in the liver and facilitate emulsification and adsorptions of lipids. BAs modify the composition of gut microbiota via AMPs and immune regulation. Conversely, gut microbiota convert primary BAs via bile salt hydrolysis and dehydroxylation to secondary BAs. BAs in the intestinal lumen signal to the CNS via direct pathways (systemic circulation) and indirect pathways, mediated through farnesoid X receptor (FXR) signaling and Takeda G protein-coupled receptor-GLP-1 (TGR5-GLP-1) signaling⁴⁵. BAs disrupted BBB gap junction proteins (occludin) and increased BBB permeability²¹². A subset of EECs (L-cells) produce GLP-1 upon the activation of TGR5 by microbiota-regulated BAs. GLP-1 can be sensed by GLP-1 receptors expressed by afferent vagal fibers²¹³, and is involved in glucose metabolism and energy homeostasis⁴⁵. BAs represent a complex interplay of microbiota-host metabolism that influences the MGBA signaling.

BAs In Stroke:

BAs play a major role in cholesterol homeostasis, and are linked to development of atherosclerosis, a major risk factor for stroke. A recent study demonstrated that direct activation of CNS TGR5 signaling counteracts diet-induced obesity, whereas genetic downregulation of hypothalamic TGR5 promotes it²¹⁴. BAs also regulate inflammation and metabolic disorders. For example, BAs inhibit NLRP3 inflammasome activation (a major host defense mechanism) via the TGR5-cAMP-PKA axis²¹⁵. A single-center retrospective study of 777 AIS patients divided into four groups based on the quartile of the serum total BAs showed that increased admission serum BAs was associated with a reduced 3-month mortality after stroke²¹⁶, suggesting protective effects of bile acids in ischemic stroke patients. Interestingly, patients with the highest serum BAs from the same study had the lowest WBC counts, hinting at potential anti-inflammatory effects of BAs. Another study showed that lower BA excretion in feces may be an independent risk factor for stroke²¹⁷. It is challenging to reliably deduce mechanistic insights from high level correlative clinical data from different sample types (serum vs. feces). However, these retrospective studies clearly implicate the microbiota in BAs homeostasis, and BA-mediated regulation of multiple stroke risk factors. Future studies are necessary to determine the effects of age-related changes in the gut microbiota and stroke-induced dysbiosis on BAs metabolism in the context of stroke risk and recovery.

Bacteria That Regulate Tryptophan Metabolites:

Host metabolism of tryptophan is through the kynurenine pathway via tryptophan 2,3-dioxygenase and indoleamine 2,3-dioxygenase (TDO and IDO) as well as the 5HT pathway via tryptophan hydroxylase. Microbial metabolism of tryptophan is through the indole pathway by both tryptophanase-positive and tryptophanase-negative bacteria⁶². Microbiota-derived indole-based metabolites and host-derived kynurenine-based metabolites are physiologically relevant endogenous ligands for the AHR. For example, activation of AHR promotes differentiation of naïve T cells to anti-inflammatory T_reg cells²¹⁸, and activation of AHR by microbiota-dependent indole-based ligands reduces astrocyte-mediated neuroinflammation in EAE⁷⁴. Tryptophan is also a precursor for 5HT synthesis, which can modulate the signaling between the host’s EECs and ANS, as described above²¹⁹. Future studies are needed to examine the role of microbiota-dependent tryptophan-based ligands of AHR in post-stroke neuroinflammation.

Aging is associated with dysregulated amino acid metabolism, a contributor to immunosenescence and inflammaging¹⁰¹. Multiple amino acids, including glutamine, arginine, and tryptophan, have been linked to age-related remodeling of immunity¹⁰¹. Tryptophan is metabolized by both the host (kynurenine and serotonin pathways) and the microbiota (indole pathway). Aging is associated with increased IDO/TDO-mediated breakdown of tryptophan into kynurenine metabolites and leads to age-dependent elevated serum kynurenine. Metabolic shunting of tryptophan away from the indole pathway (regulated by the microbiota) toward the kynurenine pathway (regulated by the host) has been linked to the post-stroke inflammatory response^220,221. Cuartero et al. showed that activation of AHR by kynurenine exacerbates acute brain damage following ischemia²²². Pharmacological inhibition of the AHR after stroke reduced brain levels of inflammatory cytokines (IL-1β, IL-6, IFN-g, CXCL1, and S100b) and improved stroke outcomes^222,223, suggesting that activation of AHR pathway by host-derived kynurenine metabolites plays a detrimental role in response to stroke. However, the role of AHR activation is highly ligand- and cell-specific⁷⁴, and future studies are required to elucidate the effects of microbiota-derived (indole-based) versus host-derived (kynurenine-based) ligands of AHR after stroke.

Trimethylamine N-oxide (TMAO):

Diet has profound effects on the microbiota. For example, the production of TMAO by the microbiota can be increased by a high intake of dietary precursors of choline (L-carnitine and phosphatidylcholine), found in red meat and eggs²²⁴. Microbiota-derived TMAO has been linked to increased risk of cardiovascular events^97,225–228. TMAO production involves the conversion of choline and L-carnitine by microbial lyases to TMA, an odorous gas. TMA reaches the liver via the portal vein, after which Flavin monooxygenases convert it to TMAO²²⁹. Clinical studies of plasma and urinary levels of TMAO after a dietary intake of TMAO precursors (e.g., phosphatidylcholine) in healthy participants before and after suppression of microbiota with antibiotics showed that TMAO production depends on metabolism by the microbiota²²⁸. Higher plasma levels of choline and betaine (a TMA-containing essential nutrient and its oxidation product) are associated with 1.9- and 1.4-fold increased risk of major adverse cardio/cerebrovascular events, including stroke²³⁰. Importantly, the detrimental effects of TMAO on stroke severity and outcomes can be transferred through FMT, provided that microbial lyases are present to convert choline to TMA²³¹. Another study showed that levels of proinflammatory Ly6C^high monocytes were higher in mice fed with choline-rich diet to increase TMAO synthesis, but this increase in inflammatory monocytes was abolished if mice received antibiotics with the choline-rich diet²³². FMT studies have shown that stroke severity is transmissible and the TMAO pathway may be a mediator of this transmissibility²³¹. Specifically, the microbial cutC gene (required for choline-to-TMA conversion) is sufficient to increase TMAO production and increase infarct size after stroke²³¹. The role of TMAO in cerebrovascular disorders has been reviewed²³³. Improvement of diet may have a large effect on both the host and the microbiota.